A New Method for Finite-frequency Boundary Sensitivity Kernels — Calculation and Application.
Abstract
Converted seismic waves from the seismic discontinuities at depths of about 410 km and 670 km are used frequently to study the physical and chemical properties of the mantle transition zone (MTZ). In the conventional receiver function method, ray theory, which is based on the high-frequency approximation, is used for imaging these two relevant seismic discontinuities, although the data used are typically low-pass (≥5s) filtered in practice to suppress high frequency noise and obtain a more coherent image of the discontinuity. Diffractional effects on the direct and converted seismic waves are generally neglected, leading to biased results. Recent studies have shown that ray theory overestimates traveltime anomalies particularly for small-scale (≤ 400 km) undulations of the discontinuities. Therefore, it makes sense to introduce finite-frequency sensitivity kernels into the analysis of receiver functions.
In this study, we propose a new method to calculate boundary sensitivity kernels for receiver functions considering finite frequency effects. Based on the variational principle, we derive analytic expression of the boundary sensitivity kernels in terms of synthetic Green's functions, which can be efficiently pre-calculated using the code QSSP for a spherically symmetric earth model. Here, we demonstrate the boundary sensitivity kernels for direct P wave arrival times as well as the arrival times of converted waves imaged with receiver functions. Using these sensitivity kernels, we calculate the traveltime anomalies of the P waves caused by a Gaussian undulation of the 670 km discontinuity. The predicted delays, which are validated with 3D synthetic data simulated using the Spectral Element Method, show significant differences to those predicted by the ray theory. The new approach presented here has several advantages. In particular, the use of Green's function database prepared with QSSP makes the calculation of the sensitivity kernels fast and accurate within the complete frequency band, making application of finite-frequency receiver function kernels possible even for large data sets.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2018
- Bibcode:
- 2018AGUFM.S53C0411D
- Keywords:
-
- 7260 Theory;
- SEISMOLOGYDE: 7270 Tomography;
- SEISMOLOGYDE: 7290 Computational seismology;
- SEISMOLOGY